I feel decentralized storage and better tools are the key both for technical and social reasons. We should start looking at old data just like we look at old paintings, old parchment books and fossils - as things with inherent intangible value to be preserved for, and by, every generation.

Centralized archives are single points of failures subject to political and business pressures. Political ideologies and their governments love to hide or rewrite inconvenient histories. I feel centralized archives (like Wayback Machine or Google's initiative) can't be trusted or relied upon for long-lasting access.

Something that surprises me is that the tech community seems to feel the necessity for long-term archiving much more strongly than the history / sociology / humanities research communities. I expected university humanities departments would be at the core of such initiatives.

The solution is decentralization, like https://d.tube, https://joinpeertube.org, https://www.bitchute.com or https://lbry.com.

Libya 2011 and arab spring in general spawned a lot of violent videos that were important for documenting what happened. They must have been archived or stored somewhere too - I know they were on youtube at the time, quickly taken down, and some were trying to archive/stow them.

Same as in other industries, use something else to cover the costs, best if there is incentive to keep that service alive. For instance, a porn site like xvideos could take over the hosting and moderation.

web.archive.org